Individual and spatial determinants of mortality during the Covid-19 pandemic: The case of Belgium in 2020 (preprint)

Context. The year 2020 was marked by the Covid 19 pandemic. In Belgium, it led to a doubling in deaths, mainly grouped into two periods. This article aims to compare the relative importance of predictors and individual and spatial determinants of mortality during these two waves to an equivalent non-pandemic period and to identify whether and to what extent the pandemic has altered the sociodemographic patterns of conventional mortality. Methods. The analyses relate to all-cause mortality during the two waves of Covid-19 and their equivalent in 2019. They are based on matching individual and exhaustive data from the Belgian National Register with tax and population census data. A multi-level approach was adopted combining individual and spatial determinants. Results. Mortality patterns during the pandemic are very similar to those observed outside the pandemic. As in 2019, age, sex, and household composition significantly determine the individual risk of dying, with a higher risk of death among the oldest people, men, and residents of collective households. However, their risk of death increases during the Covid period, especially in the 65/79 age group. Spatial information is no more significant in 2020 than in 2019. However, a higher risk of death is observed when the local excess mortality index or the proportions of isolated or disadvantaged people increase. Conclusions. While the Covid pandemic did not fundamentally alter conventional mortality patterns, it did amplify some of the pre-existing differences in mortality.

A world apart: levels and factors of excess mortality due to COVID-19 in care homes. The case of Wallonia - Belgium. (preprint)

COVID-19 became pandemic in 2020 and causes higher mortality in males (M) than females (F) and among older people. In some countries, like Belgium, more than half of COVID-19 confirmed or suspected deaths occurring in spring 2020 concerned residents of care homes. The high incidence in this population is certainly linked to its peculiar age structure but could also result from its poorer general health condition and/or from a higher contamination through the staff of care homes, while protection equipment and testing capacity were initially limited. To address these issues, we used data from Wallonia (Belgium) to characterize the distribution of death rates among care home institutions, to compare the dynamics of deaths in and outside care homes, and to analyse how age and sex affected COVID-19 death rates inside and outside care homes. We also used annual death rates as a proxy for the health condition of each population. We found that: (1) COVID-19 death rate per institution varied widely from 0{per thousand} to 340{per thousand} (mean 43{per thousand}) and increased both with the size of the institution (number of beds) and with the importance of medical care provided. (2) 65% of COVID-19 deaths in Wallonia concerned residents of care homes where the outbreak started after but at a faster pace than the outbreak seen in the external population. (3) The impact of age on both annual and COVID-19 mortality closely follows exponential laws (i.e. Gompertz law) but mortality was much higher for the population living in care homes where the age effect was lower (mortality rate doubling every 20 years of age increment in care homes, 6 years outside them). (4) Both within and outside care homes, the ratio of M/F death rates was 1.6 for annual mortality but reached 2.0 for COVID-19 mortality, a ratio consistent among both confirmed and suspected COVID-19 deaths. (5) When reported to the annual death rate per sex and age, the COVID-19 relative mortality was little affected by age and reached 24% (M) and 18% (F) of their respective annual rate in nursing homes, while these percentages reduced to 10% (M) and 9% (F) in homes for elderly people (with less medical assistance), and to 5% (M) and 4% (F) outside of care homes. In conclusion, a c. 130x higher COVID-19 mortality rate found in care homes compared to the outside population can be attributed to the near multiplicative combination of: (1) a 11x higher mortality due to the old age of its residents, (2) a 3.8x higher mortality due to the low average health condition of its residents, and (3) probably a 3.5x higher infection rate (1.6x in homes for elderly people) due to the transmission by its staff, a problem more acute in large institutions. Our results highlight that nursing home residents should be treated as a very specific population, both for epidemiological studies and to take preventive measures, due to their extreme vulnerability to COVID-19.